WORKING PAPER

THE ECONOMIC IMPACT OF

MERGERS AND ACQUISITIONS

ON CORPORATIONS

Working Paper Number 4February 1995

WORKING PAPER

THE ECONOMIC IMPACT OF

MERGERS AND ACQUISITIONS

ON CORPORATIONS

by Gilles Mcdougall, Micro-Economic Policy Analysis,Industry Canada.

Working Paper Number 4February 1995

Aussi disponible en français

ACKNOWLEDGEMENTSI would like to thank Jack Lothian of the Industrial Organization and FinanceDivision at Statistics Canada for all the work he did on this research project. Hecollaborated on the development of the database and the econometric analysis. Ourdiscussions were very helpful on more than one occasion. I would also like to thankJanice McMechan and Joan Farnworth, of the same Division, for their help in theinitial stages of this project.

I am also indebted to John Knubley, Someshwar Rao and Ross Preston for thefruitful exchanges I had with them. Finally, I would like to thank the three externalreaders for their very stimulating comments on a previous version of this text.

The views expressed in these working papers do not necessarily reflect those ofIndustry Canada or of the federal government.

Details of the titles available in the Research Publications Program and how toobtain copies can be found at the end of this document.

Comments should be addressed to:

Someshwar RaoDirector, Strategic Investment AnalysisMicro-Economic Policy AnalysisIndustry Canada 5th Floor, West Tower235 Queen StreetOttawa, Ontario K1A 0H5

Telephone: (613) 995-7077

TABLE OF CONTENTS

SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

OVERVIEW OF RESEARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3U.S. Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4Canadian Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

THE CONCEPTUAL FRAMEWORK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

THE DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Main Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

ECONOMETRIC ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Profit to Sales Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25The Research and Development Expenditures to Sales Equation . . . . . . . . . . . . . 26

CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

APPENDIX ADESCRIPTION AND LIMITS OF THE DATA SET . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

APPENDIX BANALYSIS OF THE PROFIT FUNCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

RESEARCH PUBLICATIONS PROGRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

i

SUMMARY

While interest in mergers and acquisitions is not new, it has intensified recently as a resultof a plethora of mergers and acquisitions in the 1980s. The importance of takeovers by foreigninterests also increased in relative importance during that period. Several studies were conductedat that time to shed light on the motives for these transactions and to determine the consequencesby evaluating the costs and benefits for both the corporations and the countries in which theywere located. In the current climate of globalization, it is not surprising that interest in themergers and acquisitions phenomenon has remained high.

The studies evaluating the profitability of corporations which were the objects of mergersor acquisitions have been using two different approaches:

The financial approach examines trends in the share prices of corporations involved inmergers or acquisitions and compares them to a reference group of corporations. Corporateperformance is considered to have improved if the returns to shareholders are greater after theacquisition or merger. The results obtained using this approach, largely in the United States butalso in Canada, show that corporate takeovers generally have favourable consequences forshareholders. Stock markets seem to take a positive view of announcements that corporations willbe merged or taken over.

The industrial organization approach examines certain financial or economicperformance variables of corporations before and after they have been taken over. Trends in thesevariables, as compared to a reference group, provide an indication of the net effect of theacquisition on profitability rates.

This study adopts the second approach. As part of Industry Canada's larger researchprogram into the overall micro-economic adjustment process of corporations, this paper hopes toanalyze the impact of mergers and acquisitions on corporate decisions and corporate performance.

The data used in this research was constructed by Statistics Canada for the specificpurpose of the study. It is based on a list of firms that were the objects of merger or acquisitionbetween mid-1985 and end of 1987. The observations are those takeovers that had to beapproved under the Investment Canada Act. Because the Act establishes a minimum threshold onthe value of the transaction to be reviewed, the firms included in the sample are of a large size. They are also, by definition, firms that become foreign-controlled after the transaction.

The usual statistical techniques were used to disentangle the impact of the takeover on theperformance of the continuing firm.

ii Summary

Some of the important findings of the research are:

• Corporations that have been taken over by foreign interests increase their capitalinvestment and their R&D spending.

• However, short-term profitability is not positively affected by foreign takeovers. In fact,the profit to sales ratio declines sharply immediately after takeovers by foreigners. Theprofit to equity ratio behaves in the same way.

• High levels of R&D spending seem to be associated with high profitability. In otherwords, a high level of investment in the production and use of new technologies is causingthe profitability of firms to increase.

• Firms appear to undertake capital investment in tandem with R&D spending, implyingcomplementarity between the two inputs in the production process.

• There exist economies of scale in the R&D activity sphere. This means that as the firm isgrowing in size, the relative amount of R&D spending it needs to do declines.

• The behaviour of corporations taken over by foreign interests differs significantly fromcorporations taken over by Canadian interests. The latter seems to show an increase inshort-term profitability and, at best, no change in investment in physical capital or researchand development.

These findings imply an interesting adjustment process to a foreign takeover. Immediatelyafter the change of ownership, corporations seem to take a longer-term perspective and invest inR&D, or physical capital. To do this, they are prepared to accept a short-term reduction inprofitability.

This research shows, in our view, the importance of analyzing the longer-termconsequences of mergers and acquisitions. The relatively poor profitability of the target firm inthe first few months and years after the transaction apparently reflects a period of adjustmentduring which the two different work cultures get to know each other and learn how to co-operate.In addition, corporations that invest in acquisitions have to take the time to recognize and use, ina profitable way, all aspects of the assets they have acquired.

Corporate assets are becoming increasingly intangible. This means that assets areincreasingly composed of ideas, knowledge and know-how that cannot be easily codified. Theremay be embodied in a piece of equipment, or in the people within the firm. The internalizationand technological competence theories suggest that acquiring these intangible assets is animportant motive in making acquisitions. To profit fully from these intangible assets, corporationsneed time and complementary investments in physical capital and research and development. Thefindings of our research suggest that these two theories might lead to a correct interpretation ofthe motives for mergers and acquisitions activities.

1

INTRODUCTION

While interest in mergers and acquisitions is not new, it has intensified recently as a resultof a plethora of mergers and acquisitions in the 1980s. Several studies were conducted at thattime to determine the effects on Canada of takeovers by foreign interests or by other Canadiancorporations, and interest in this area has remained high — which is not surprising given thecurrent climate of globalization.

Despite the intensity of this interest, there is a lack of consensus on the effects of mergersand acquisitions on the countries in which they take place and on the targeted corporations. Manystudies have attempted to analyze and quantify the costs and benefits of mergers and acquisitions.However, the conclusions reached in these studies are so diverse that it is impossible to arrive at aclear, unequivocal consensus of opinion. In a working paper (No. 13) published by InvestmentCanada, Mr. Shapiro attributes this diversity of views to the multitude of hypotheses advanced byresearchers and, most importantly, to the poor quality of the databases, which make it impossibleto determine which of the various hypotheses are more accurate.

Using an approach similar to that used in the other studies, this paper attempts todetermine the costs and benefits of mergers and acquisitions. It is a follow-up to InvestmentCanada's Working Paper No. 11, Business Performance Following a Takeover. It uses howevera different data base and obtains results from a different methodology.

The economic models and econometric techniques used in this paper reflect the approachnormally taken in studies of this nature, and are an improvement over the model used in WorkingPaper No. 11. Moreover, the database is original. It was designed and constructed specifically forthis research.

In the next section, we describe aspects of the published research on the economicevaluation of mergers and acquisitions. We then define the conceptual framework within whichwe will evaluate the effects of mergers and acquisitions, present the methodology used to developthe database and describe the main features of the data. Finally, we present some of the results ofour econometric analysis.

3

The internalization theory is based on the idea ofintangible assets: in order for corporations to attractmergers or acquisitions, they must have intangibleassets that make them profitable. These assets caninclude knowledge of a particular market, know-howin a particular technology or an enviable reputation forproduct quality. Usually, these assets have two majorcharacteristics: they must have the attributes of apublic good (i.e., their running costs within thecorporation must be zero) and they must have hightransaction costs so the most profitable way ofacquiring them is through mergers or acquisitionsrather than purchase or rental.

The internalization theory assumes that the purchasersof the targeted corporations want to obtain theirintangible assets. These assets will produce acompetitive advantage that should eventually findexpression in increased profit. When intangible assetsare recognized, it is usually by competitors. As aresult, the internalization theory best explainshorizontal mergers.

Box 1:Internalization Theory

OVERVIEW OF RESEARCH

Evaluating the performance of corporations involved in mergers or acquisitions has beenthe subject of a great deal of research. Large, and sometimes spectacular, mergers andacquisitions have attracted media coverage that has stimulated the interest of both researchers andthe general public. Attempts were made to shed light on the motives behind these transactions andto determine their consequences by evaluating the costs and benefits for both the corporations andthe countries in which they were located.

The variety of reasons for mergers and acquisitions and the diversity of their consequenceshave given rise to a range of hypotheses, eachof which attempts to explain part of thisphenomenon. These hypotheses can besubsumed into three major theories:

• internalization theory;• technological competence theory;• transaction cost theory.

The internalization theory describedin Box 1 suggests that corporations attemptto acquire others because they want toprocure intangible assets that generally givethem a competitive advantage.

Another theory, derived from thefirst, is the technological competence theoryexplained in Box 2. According to this theory,corporations that engage in mergers andacquisitions are attempting to "internalize"technological advantages by acquiring thecorporations that possess them.

A third theory, that of transactioncosts, applies to vertical mergers andacquisitions aimed at reducing uncertainty orthe cost of procuring particular factors ofproduction. It is explained in Box 3.

4 Overview of Research

A corporation may decide to acquire animportant supplier in order to ensure that aparticular input is available, to reduce supplyuncertainties or to reduce the cost of this input.This is the basis of the transaction cost theorywhich applies primarily to vertical transactions.It is more likely to come into play if the numberof buyers and sellers in the market are limited,information about costs and prices is limited andthe cost of changing suppliers is not negligible.

Box 3:Transaction Cost Theory

The mounting importance of technology has given rise to a new theory that is an extension of the internalizationtheory, namely the technological competence theory recently developed by John Cantwell and based on theinternalization of intangible technological assets. It assumes that technology consists of two factors: one that canbe codified (e.g., written information about the technology, plans, etc.) and another that cannot be codified (e.g.,certain abilities needed to operate it, particular knowledge, the ways in which it operates, etc.). It is this latterfactor that constitutes technological competence, that is to say, an intangible asset. Technological competence isthought, moreover, to be of cardinal importance for corporate success.

This theory has certain consequences. First, when targeted corporations are in industries with high technologicalcoefficients, potential purchasers will be more inclined to install research and development capacity there, thusenhancing local innovation. Second, when local corporations have low technological capacities, mergers andacquisitions may increase the technological content of production. Third, in intermediary cases when corporationsengage in research but are not on the cutting edge of technology, mergers and acquisitions may result in thecomplete absorption of the targeted industry.

Box 2:Technological Competence Theory

These three theories show the variation inreasons for a merger or acquisition. Khemani(1991) expressed this very well:

There are multiple reasons, motives,economic forces and institutional factorsthat can, taken together or in isolation,influence corporate decisions to engagein mergers or acquisitions. Over the lastfew years, the pressures emanating frominternational competition, financialinnovation, economic growth andexpansion, heightened political andeconomic integration, and technologicalchange have all contributed to the increased pace of mergers and acquisitions. Of course, mergers and acquisitionscan still be motivated by such classic commercial and economic considerations asbroadening the range of related products and the geographic market,diversification, and the risks and benefits of vertical integration. Finally, new ormodified tax regimes, the cost of capital, and policy on such things as foreignproperty, economic regulations and privatization also have an effect on theintersectoral/international variations in the number of mergers and acquisitions.

Overview of Research 5

It can be assumed that these reasons and motivations have enhanced corporate profitabilityas the ultimate, long-term objective. It seems reasonable to assume that, even if this is not alwaysthe case, the ultimate concern of corporate managers who make acquisitions, regardless of theirmotives at the outset, is increasing long-term profit. However, this is affected by so many otherfactors that it can become very difficult to make isolated statistical measurements of the effect ofmergers or acquisitions on profit.

The "free cash flow" theory developed by Jensen (1988) provides a good example ofintermediate objectives that can lead to greater profitability in the long run. This theory assumesthat corporate shareholders do not necessarily share the same objectives as the managers. Theconflicts between these differing objectives may well intensify when corporations are profitableenough to generate "free cash flow," i.e., profit that cannot be profitably re-invested in thecorporations. Under these circumstances, the corporations may decide to make acquisitions inorder to use these liquidities. These acquisitions are often financed both by issuing debentures andliquidating the cash in hand. It is therefore higher debt levels that induce managers to take newmeasures to increase the efficiency of corporate operations. According to Jensen, long-term profitcomes from the re-organization and restructuring made necessary by takeovers.

Studies that attempt to evaluate the profitability of corporations that have been the objectof mergers or acquisitions can be categorized according to whether they take a financial orindustrial organization approach.

The financial approach examines trends in the share prices of corporations involved inmergers or acquisitions and compares them with a reference group of corporations. Corporateperformance is considered to have improved if the returns to shareholders are greater after theacquisition or merger. The results obtained using this approach, largely in the United States butalso in Canada, show that corporate takeovers generally have favourable consequences forshareholders. Stock markets also seem to take a positive view of merger or acquisitionannouncements.

The industrial organization approach looks at certain financial or economicperformance variables of corporations before and after they have been taken over. Trends in thesevariables compared with a reference group provide an indication of the net effect of theacquisition on profitability rates. Since the current study generally follows this second approach,we will analyze it in more depth. Boxes 4 and 5 provide summaries, in the form of tables, of themain conclusions that have been reached.

U.S. Studies

In the United States, some studies that took this approach concluded that the performanceof acquired corporations remains, at best, unchanged and usually declines. For instance,Ravenscraft and Scherer (1987 and 1989) showed that the profitability of

6 Overview of Research

Authors Subject of the Study Results

Ravenscraft and Scherer(1989)

- Ex ante analysis of the profitability ofacquisitions for the targeted corporations- Ex post analysis of operating results- Manufacturing sector in the U.S.- 1957-77

- The targeted corporations (1989) were veryprofitable ex ante- Ex post profitability of the targetedcorporations decreased sharply

Caves(1989)

- Survey of the published studies- Use of ex ante case studies and ex postevaluation- U.S.

- The ex ante case studies showed very slightincreases in the value of the stock of thetargeted corporations and the purchasers

Lichtenberg(1992)

- The relationship between changes in controlof companies, productivity and investment inresearch and development- U.S.- 1972-81

- Large increase in total productivity of thefactors of production after takeovers

Brown and Medoff(1988)

- The effect of acquisitions on corporatewages and employment- U.S.

- Mergers were associated with approximatelya four percent decrease in wages and a twopercent increase in total employment

Hall(1988)

- The effects of acquisitions on investment inresearch and development- Manufacturing sector in the U.S.- 1976-85

- No indication that acquisitions result in areduction in research and developmentexpenditures- Corporations that successfully innovate arethe preferred targets of potential purchasers

Box 4:Summary of A Few Important Studies in the United States

corporations decreases after takeovers. Similarly, Caves (1989) concluded that the market shareand productivity of acquired corporations declined faster than that of corporations that had notbeen take over.

On the other hand, Lichtenberg (1992) concluded that corporate efficiency improved aftertakeovers. He examined changes in the total productivity of the factors of production during theseven years before and after takeovers in the manufacturing sector and observed that immediatelybefore takeovers, the total productivity of the factors of production of the targeted corporationswas significantly lower than that of other corporations. However, this gap narrowed considerablyover time with the result that seven years after the takeovers, the differences in productivitybetween firms that had been taken over and those that had not were insignificant. These increasesin productivity were partly due to a reduction in total employment. Lichtenberg also studiedemployment that was expressly related to research and development and concluded that therewere no significant differences. Finally, his results confirmed that the market share of corporationsthat have been taken over declines after the acquisition is made.

Overview of Research 7

Authors Subject of the Study Results

Baldwin and Gorecki(1987)

- Analyzes the relationship betweenacquisitions and productivity- Canadian manufacturing sector- 1971-79

- Takeovers are associated with strongincreases in productivity

Baldwin and Caves(1990)

- Examines the effect of takeovers on marketshare and productivity- Provides data according to the amount offoreign control- Canadian manufacturing sector- 1970-79

- Takeovers result in increases in market shareand productivity- The favourable effects noted above increasewith the amount of foreign control

Tarasofsky and Corvari(1991)

- Analyzes the relationship betweenprofitability and mergers and acquisitions- Canadian manufacturing sector- 1983-87

- Canadian takeovers do not causeunprofitable use of company assets- Profits remain unchanged after takeovers

Allen(1992)

- Analyzes the performance of companies thatwere taken over - Canadian manufacturing sector - 1983-87

- The economic performance of companiesimproves after takeovers- The amount of investment in physical capitaland research and development exceeds thesectoral average

Box 5:Summary of a Few Important Studies in Canada

Brown and Medoff (1988) also examined the effect of acquisitions on the work force.They concluded that the public's general impression that acquisitions have a negative effect onemployment is not true. They studied 200,000 corporations between 1978 and 1984 and foundthat mergers were associated with about a two percent increase in total employment and about afour percent decrease in wages.

Hall (1988) studied the effect of takeovers on research and development. He examined arelatively limited group of corporations in the manufacturing sector between 1976 and 1985.According to him, corporations that successfully innovate are more often the target of takeoverattempts. However, the overall research and development expenditures of corporations that hadbeen taken over did not differ significantly from corporations that had not been taken over.

8 Overview of Research

Canadian Studies

In Canada, the various studies fail to reach a consensus on the consequences of takeovers.To analyze determining factors, Baldwin and Gorecki (1987) studied the process through whichcorporations enter the Canadian manufacturing sector. They examined two differing entryprocesses: acquisitions and the establishment of new plants. They also examined the differencesbetween Canadian-owned and foreign-owned corporations. Although Baldwin and Gorecki didnot set out directly to explain the effect of mergers on profitability, they shed some light on thereasons why corporations enter particular industries. They concluded that high profit expectationsare more associated with acquisitions than with the establishment of new corporations. However,they only found this to be true for acquisitions by Canadian corporations. Foreign corporationsseemed less concerned about profitability. Regardless of whether they entered an industry bycreating new plants or through acquisitions, foreign corporations did not react significantly to thegrowth and profitability variables to which their local competitors seemed so sensitive. Theseresults jibe with the conclusion reached later in this analysis that foreign corporations aremotivated to enter industries more by global than by local considerations. We will see below thatthese "more global considerations" of foreign corporations play an important part in explainingthe results that we obtained.

In a more recent study, Baldwin and Caves (1990) examined trends in added value perworker between 1970 and 1979 in corporations that had been the object of mergers or takeoversduring that period. They concluded that changes in control were followed by improvements inproductivity. Baldwin and Caves also attempted to determine if the presence of a large number offoreign corporations in a particular industry changed the observed results for all industries. Theydivided their sample into three groups depending on the proportion of shipments made bycorporations under foreign control. They concluded that the productivity gains followingtakeovers were larger in industries with high levels of foreign control, i.e., in industries withshipments majority-controlled by foreign corporations.

The above results are especially interesting when compared with each other. They showhow important the selection of performance indicators is in judging mergers and acquisitions.However, they also show how important it is to judge the effects of mergers and acquisitions in abroad, evolving context. Baldwin and Gorecki (1986) pointed out that foreign-controlledcorporations do not concern themselves much with the profitability of their mergers andacquisitions, at least in the short run. On the other hand, Baldwin and Caves (1990) claimed thatlabour productivity increases after takeovers, especially if these takeovers are by foreigncorporations.

It is essential to have enough information to adequately judge these increases inproductivity. Are they the result of significant reductions in employment? Are they accompaniedby corresponding increases in real wages in the corporations? These are important questions in anoverall analysis of mergers and acquisitions. The increases in productivity after takeovers shouldeventually result in greater profitability in the foreign-owned corporations. This could mean that

Overview of Research 9

foreign-owned corporations tend to take a longer-term perspective when evaluating mergers andacquisitions.

Tarasofsky and Corvari (1991) obtained results that were the opposite of those describedabove. They believe that corporate profitability remains, at best, unchanged after takeovers and, inmany cases, actually diminishes. Tarasofsky and Corvari used several different measures ofcorporate profitability. One was based on stock prices and constituted a measurement of the profitdistributed to shareholders. Two other variables were based on accounting measurements ofprofitability, namely the rate of return on equity and the rate of return on assets. Both measures ofprofitability produced the same conclusion: takeovers improved profitability in only about40 percent of cases. In the other cases, profitability remained, at best, unchanged, but usuallydeclined.

The research in Canada and the United States is therefore rather ambiguous about theconsequences of mergers and acquisitions. The results are crucially dependent on themeasurement of profitability that is chosen, the conceptual framework and the methodology usedfor the analysis. The following section looks at the conceptual framework used for analyzing theresults that follow.

There is another current in the research based on a different objective than maximizing profits. It assumes that corporate1

financial transactions aim to maximize the interests of the managers making the acquisitions. Although this research producessome interesting results, we do not emphasize it here.

10

THE CONCEPTUAL FRAMEWORK

Reasons for mergers and acquisitions are numerous and include:

• to diversify or expand markets;• to acquire particular production technologies;• to take advantage of work forces with particular skills; or• to benefit from "good opportunities" to take over a corporation.

These motives are ultimately related to a common objective: maximizing profit or returns forshareholders.1

For the purposes of this research, we assume that all corporate mergers or acquisitions areultimately motivated by profit maximization. This implies that the managers involved in mergersand acquisitions take necessary and adequate measures to increase the efficiency of theircorporate operations. However, these measures can be as varied as they are indirect, and it isimportant to be able to assess their effectiveness in an evolving environment of profitmaximization.

Profit for a given period (month, year, etc.) is defined simply as the difference betweentotal revenues from the sale of a company's goods and services and the costs incurred to purchasethe factors of production needed to produce these goods and services over the period in question.In reality, there are many different factors of production, but for the purposes of this study, weassume that there are a limited number: labour, physical capital, technological capital, humancapital and the capacity utilization rate.

The inclusion of the capacity utilization rate may seem surprising. In response to changesin their environment, corporations may decide to change their production levels without alteringtheir demand for other factors. Suppose, for example, that the demand for a product declines. Ifthe corporations that supply this product expect the decline to be temporary, they may decide toreduce their capacity utilization rate. This implies under-utilization of all factors of production,including labour. However, this option, which reduces profits in the short-run, could provebeneficial in the long-run. Had the corporation reacted, instead, by cutting jobs, the costsassociated with dismissing people and then re-hiring could have been higher than the costsassociated with the underutilization of factors. This process would thus generate losses whichcould be higher than the losses associated with reducing the capacity utilization rate. It thereforeappears possible and, in some cases, profitable for corporations to change their capacity utilizationrates rather than their demand for other factors of production.

The Conceptual Framework 11

In order to produce the goods and services they sell, corporations use labour and variousstocks of capital in specific proportions. Following a takeover, corporations may decide to changetheir utilization of various factors of production. For instance, when corporations attempt to re-organize or increase efficiency, they often decide to reduce the number of jobs. This has animmediate impact on productivity and profit.

However, the use of some other factors of production, considered fixed in the short run,can also be modified without having a strong, immediate impact on profit (other than throughdirect investment costs). Such factors are physical, technological and human capital. In changingtheir production methods, corporations may decide to change the proportion of each of thesecapital stocks. They could increase their capital/output ratio by increasing gross fixed capital; theycould attempt to strengthen their technological advantages by increasing their investments inresearch and development; or they could change the skill mix in their work force by increasingtheir investments in education and job training.

These measures are aimed at changing production processes in order to improveprofitability. They only take effect gradually over the long run. In the short run, profit is notnecessarily positively affected by these new investments, since only current expenditures areincluded in calculating profit. Corporations that are restructuring might thus accept a temporary,short-term reduction in profit. That is not to say that the takeover is not profitable. As we sawabove, corporations may attempt to increase their long-term profitability by taking a series ofmeasures to change their factors of production, which are considered fixed in the short-run. Inthis case, the short-term decline in profit is only temporary, and the long-term profit increase ispresumably strong enough to compensate for the short-term decline.

In this case, ideally we would be able to determine a corporate profit function or,amounting to the same thing, a cost function in a system that also includes equations for theshares of the factors of production. However, to estimate this equation system, it is necessary tohave a large number of variables such as product prices and the prices of all the factors ofproduction, as well as the quantities of the factors, such as the work force employed, physicalcapital stocks, research and development expenditures, and investment in human resources. Inorder to be able to estimate a corporation's adjustment process after a takeover, longitudinal dataon both corporations would also be needed. Adaptation occurs over a long period, and one wouldneed to be able to observe such variables as profit and the productivity of the factors ofproduction for several years after the takeover.

Unfortunately, the information currently in data banks about corporations involved intakeovers does not make it easy to analyze mergers and acquisitions in such an aggregate,dynamic way. Statistics Canada is developing a new data bank that would merge several sourcesof data. It would be longitudinal and would include several of the variables mentioned above.However, this data bank is not yet developed to a point where it can be used for research.

(A

S)i ' "

0% "

1× Si % "

2× (

RDS)i % "

3× ri % "

4

12 The Conceptual Framework

See Appendix B for a more conceptual presentation of the profit function used.2

The data bank we used is therefore incomplete. Productivity, wages and employment, forinstance, were not available. In addition, the time dimension was lacking. We only had access toonly two observation points: before and after the takeovers. Under these circumstances, themodel we present here is quite limited in comparison with the "ideal" version. It includes only twoequations: a profit equation and an investment in research and development equation. Theseequations cannot capture the dynamic process that corporations go through in adjusting totakeovers. They serve, however, to compare the situation before and after the takeovers.Unfortunately, the structure of these equations is determined more by data availability than by arigorous underlying theoretical framework. The results are nevertheless useful because theyenable us to check the "statistical" validity of the preceding statements.

The first equation is for corporate profit. A profit function is normally expressed in valuespace, that is to say, with the price of inputs and products as arguments. However, these variablesare not all available in our database. We will therefore have to rely on a mixed space of quantitiesand prices. The profit equation looks as follows: 2

(1)

whereA represents the profit of corporation ii

S total sales of corporation ii

RD the research and development expenditures of corporation ii

r the real interest rate in corporation i's industryi

U the unemployment rate in corporation i's industryi

D a dichotomous variable equal to 1 if the company was the object of a merger ori

acquisition or to 0 otherwise" the coefficients to be determined for i = 1 to 5.i

The purpose of this equation is to analyze the profitability of corporations that were takenover. It is therefore important to use profitability measurements as they are perceived bycorporate managers, that is to say, those who make the economic decisions. For this reason, wepreferred not to use such measurements as the profit to equity ratio, which is more a concern ofshareholders and investors. Two possibilities remained: the ratio of profit to assets or of profit tosales. The ratio of profit to assets depends on corporate production structures, and inter-corporate comparisons are often difficult. Thus, we selected the profit to sales ratio as thedependent variable.

The total sales variable "S" is included in the equation to capture the effects of scale oncorporate profit rates. Its coefficient could be positive or negative depending on whether returns

(RDS)i ' $

0% $

1× Si % $

2× (

A

S)i % $

3× (

KS)i %

The Conceptual Framework 13

are increasing or decreasing. Some of the cyclical influences might also be captured by thisvariable.

The equation also includes a variable for the amount of expenditures on research anddevelopment. One would expect the coefficient for this variable to be positive, for there isgenerally a positive connection between corporate investment in research and development andprofitability, at least in the long run. Ideally, therefore the equation should include, for estimationpurposes, a variable for the amount of research and development lagged by several periods.However, this research and development variable was only available for one period in ourdatabase. We therefore assume that current research and development expenditures aresufficiently correlated with past expenditures to enable us to establish a causal connectionbetween them and profitability.

The interest rate variable represents the cost of capital. This is a real interest rate for aspecific industry, i.e., we subtract, from the nominal interest rate, the expected inflation rate in theindustry to which the corporation belongs. The real interest rate is, of course, the one that mattersin making decisions about physical capital. Similarly, the measurement of profitability is affectedby real interest rates, not nominal rates. This variable therefore captures the effects of both thecost of capital and higher product prices on profit.

Finally, the unemployment rate variable is included to capture the possible effects of thebusiness cycle on the industry involved. One would expect this variable to have a negative signbecause employment generally increases with economic growth. Under normal circumstances, theunemployment rate declines therefore and profit rises.

On the whole, this equation is not fundamentally different from those used in variousstudies in the past. While many studies do not have a real profit equation, Ravenscraft and Sherer(1987), for example, developed an equation using various measures of profitability, including theprofit to sales ratio. This equation is analytically very similar to ours.

The second equation is also in a mixed space of quantities and values. This function issimilar to an equation of demand for capital or investment. It stands for a particular type ofinvestment, namely investment in research and development:

(2)

where $ represents the coefficients to determined for i = 1 to 6.i

This equation includes a variable for the corporation's total sales. It serves to capture thepossible effects of economies or diseconomies of scale on research and development expenditures.Most studies we consulted concluded that there are diseconomies of scale. This means thatincreases in sales are associated with less than proportional increases in research and development

14 The Conceptual Framework

expenditures. In other words, one would expect this variable to have a negative coefficient whenrelated to company sales.

The profit variable captures the possible effects of liquidity on investment in research anddevelopment. Higher profit may eventually relieve corporations of their financial constraints andtherefore stimulate research and development investment. The presence of this variable in theequation means, however, that the two equations form a simultaneous system. We have alreadymentioned the possibility that a company's profitability is positively affected by higherexpenditures on research and development. We are now considering the possibility that higherprofit itself is the reason for higher expenditures on research and development. The estimation,which should take this simultaneity into account, will enable us to highlight the relativeimportance of this dual influence.

The equation also includes a variable for the amount of capital. The presence of thisvariable is intended to capture the interdependence effect between the factors of productionconsisting of physical capital and technological capital. The "theoretically correct" variable tocapture this substitution or complementary effect is the price of physical capital. A positive (ornegative) relation between the demand for technological capital (approximated by research anddevelopment expenditures) and the price of physical capital would mean that these two kinds ofcapital are substitutes (or complementary). Unfortunately, our database does not include anadequate variable for the price of physical capital. We therefore substituted the amount of physicalcapital per unit of sales for each company. A positive coefficient would indicate a complementaryrelationship between these two forms of capital and a negative coefficient would indicate asubstitution relationship.

Real interest rates are an important element of the opportunity cost of the funds directedtoward research and development. We assume therefore that it provides a useful approximation ofthat cost in the R&D investment equation. This variable would be expected to have a negativecoefficient.

Finally, the equation also includes an unemployment rate variable. As in the profitequation, this variable serves to capture a possible cyclical effect on the amount of research anddevelopment expenditures. It is likely that at least part of the research and developmentexpenditures are influenced by cyclical factors. In a period of recovery, for example, when generaleconomic conditions are improving and predictions about the direction of the economy areoptimistic, research and development expenditures would be expected to increase, and thecoefficient would have a negative sign.

For more details, see the statistical appendix.3

17

THE DATA

Definitions3

The database used for this study was developed by Statistics Canada from raw datasupplied by Investment Canada on some 400 corporations that had been involved in mergers oracquisitions. This list included the names of Canadian corporations targeted by foreign interests inacquisition requests to Investment Canada. To be included in the sample, the acquisition requesthad to be accepted by Investment Canada between July 1, 1985 and December 31, 1987.

For that list, Statistics Canada first identified those acquisitions that targeted entirecorporations rather than divisions. The purpose was to develop a database composed solely ofentire corporations. Corporations that continued to exist after the mergers or acquisitions werealso identified. It was preferable for the group of corporations to remain the same throughout theexercise.

After this filtering process, Statistics Canada correlated the list of selected corporationswith the corporations in the data bank created from the share-capital property declarations ofcorporations subject to the Corporations and Labour Unions Returns Act (CALURA) and fromthe financial information that all corporations must provide with their tax returns. This data bankcontains some financial variables. We used assets, sales, equity and profit. We also used threeratios: profit to sales, assets to sales and profit to equity. Total research and developmentexpenditures and the ratio of research and development expenditures to total sales came fromanother data bank.

Statistics Canada measured all the variables mentioned above for two distinct periods:before and after acquisition. In all cases, 1984 was taken to be the year before the acquisition.This means that for those mergers that took place early, i.e., in July or August 1985, the intervalwas very short between the acquisition and the time when pre-acquisition performance wasmeasured. However, for the rest of the sample, this interval was about two years on average. Theyear 1988 was taken to be the year after the acquisition. This means, on average, that nearly twoyears had passed after the merger or acquisition was approved.

Taking just one year for the period before and after the acquisitions is certainly a weaknessin our database. Generally, corporate profit rates are sufficiently variable that using just oneobservation for each period may well limit how representative it is. A number of the studies welooked at used average observations over, for example, five years before and five years after thetakeover. We neither could nor wanted to develop such a database for several reasons. First, ourobjective was to analyze the behaviour of corporations that kept their own identities after the

18 The Data

takeovers. If we had taken corporations that kept their own identities both five years before andfive years after the takeovers, our sample would have been much smaller. In addition, when thisdatabase was developed, the only available financial variables were for 1988.

This study aimed to analyze the ongoing adjustment process that targeted corporations gothrough after takeovers. It was therefore important to be able to examine the behaviour offinancial variables for several years after the takeovers in order to follow developments in thesecorporations. In that respect, the immediate impact of the takeovers and the repercussions overthe next few years were equally important. The results of this study must therefore be analyzed inthis long-term framework. While future research should obtain these data for other years after theacquisition, analysis of the immediate impact would remain crucial.

The use of single observation points for the periods before and after the acquisitions madeit impossible to allow for changes over time in such variables as profitability. However, thisinability to take changes into account over time in financial variables was offset by the fact that wehad two reference groups of corporations to compare with the targeted corporations. It wastherefore possible for us to take inter-industry variations into account.

As a result of the relatively limited number of corporations involved, we had to break upsome industries in a special way. To preserve the confidential nature of the data, particularindustries were grouped together. The industrial breakdown that we used covered five sectors.

TABLE AIndustrial Representativeness of the Initial Sample

Industry Percentage of Number of CorporationsCorporations

Mines 9 12

Manufacturing and construction 51 67

Wholesale and retail trade 22 29

Finance, insurance and real estate 11 14

Services and public amenities 7 9

Total 100 131

Table A shows the percentage of corporations in each industrial group, as well as thenumber of corporations according to the industry to which they belonged at the time when theacquisition requests were made. The combined manufacturing and construction sector made upmore than half of the corporations in the sample and were over represented in comparison withtheir contribution to the total production. Twenty-two percent of the corporations were in the

The Data 19

trade sector, eleven percent in the financial sector and eight percent in the service sector. Thelatter three sectors were under represented in comparison with their contributions to totalproduction.

The list of firms thereby obtained thus forms our first group, identified hereafter as "MA". It consists of corporations that had been the objects of mergers or acquisitions. After thetakeover, they were all under foreign control. We then wanted to create a reference group ofcorporations that had not been the object of mergers or acquisitions during this period and towhich the MA group could be compared. Statistics Canada developed two of these groups. Theywere constructed to have the same industrial and size compositions. It was therefore possible tomake comparisons between the groups, "holding constant" the size of the corporations and theindustrial structure.

MA Group CC Group FC Group

- Industrial representativeness - Industrial representativeness - Industrial representativenesscorresponds to the sample corresponds to the sample corresponds to the sample

- Representativeness of the size - Representativeness of the size - Representativeness of the sizeof the corporations in the of the corporations in the of the corporations in thesample sample sample

- Canadian or foreign control - Canadian control - Foreign controlbecoming foreign control

- Change in control between - No change in control - No change in control between1985 and 1987 between 1985 and 1987 1985 and 1987

Box 6:Table of the Corporate Groups in the Sample

The CC group consisted of Canadian-controlled corporations, and the FC group ofcorporations was solely under foreign control. Comparing the MA group with the FC grouptherefore isolated the merger or acquisition effect, since ultimate control was foreign in bothcases. Comparing the MA and CC groups measured the combined effect of a takeover and adifferent country of control. Finally, comparing the CC and FC groups was also interestingbecause it measured the effect of a different control location on the behaviour of corporations thathad not been involved in mergers or acquisitions. (In Box 6 we show the differences betweenthese three corporate groups in table form.)

Main Characteristics

20 The Data

To put this in context, the average revenues of the thousand largest companies was nearly $118 million.4

In the next section, we examine the data in order to determine their main characteristics.The data are shown in Tables 1 to 8 (at the end of this report) and include the averages that weredetermined for each group (MA, CC and FC) covered in the database. The tables also include, forreference, the averages for all non-financial corporations according to whether they wereCanadian- or foreign-controlled. This data comes from Statistics Canada's CALURA data bank.

It is interesting, first of all, to see our sample in the context of the entire economy. Tables1 to 4 show that the corporations in our sample were among the largest in Canada. Table 1shows, for example, that the revenues of the corporations in our database were considerablyhigher than the revenues for all corporations. While corporations in the MA group had averagerevenues of about $131 million, the average for all Canadian-controlled corporations was only$1.4 million. These orders of magnitude remained the same when other variables were considered4

such as profit, assets or equity (see Tables 2 to 4). It is important to keep this in mind whenanalyzing the various results. Although the results are presented and analyzed in relative terms,the possibility of economies of scale could make the differences in the average sizes of thecorporations a significant factor.

Let us look now at the data in our sample. The average revenues of corporations in theMA group increased only slightly between 1984 and 1988, namely 8.6 percent. The revenues ofCanadian-controlled corporations that had not been involved in mergers or acquisitions (CCGroup) increased by nearly 32 percent, and those of foreign-controlled corporations that had notbeen involved in mergers or acquisitions (FC Group) increased by 12 percent.

Table 10 provides detailed information about revenue trends by industry. It shows thatslower revenue growth in the MA group, in comparison with the CC and FC groups, exists in allsectors, with the exception of finance, insurance and real estate.

Profits remained constant in the MA group, increased considerably in the CC group anddeclined considerably in the FC group. Three sectors in the MA group experienced increases thatwere quite large, namely manufacturing and construction, commerce and finance. The mining andservice sectors, on the other hand, experienced decreases. The profits of corporations in the CCand FC groups experienced similar trends (see Table 11).

The growth in assets between 1984 and 1988, shown in Table 3, was about the same forthe three groups of corporations, although it was a little stronger in the MA group. The detailedinformation on an industry basis in Table 12 shows quite a large increase in assets in the financial,services, and manufacturing and construction sectors.

Finally, the equity growth in the MA group fell between that of Canadian-controlled (CC)and foreign-controlled (FC) corporations. The mining and manufacturing sectors experienced

The Data 21

greater increases in the MA group than in the CC and FC groups, while the commerce andservices sectors experienced smaller increases.

Tables 5 to 8 show trends in financial ratios rather than the levels of the variablesthemselves. It should be noted that these ratios were determined by taking the average of theratios for each corporation in the sample, not by taking the ratio of the average levels shown inTables 1 to 4. Table 5 and Graph 1 show, first, that there was a general decrease in profitabilitybetween 1984 and 1988: the ratio of profit to sales in the three groups decreased considerablyduring this period. Although the profit margins were positive in 1984, they became negative forthe MA and CC groups and about zero for the FC group. Profit margins decreased the most inthe group of corporations that had been the object of mergers or acquisitions by foreigncorporations.

Figure 1Ratio of Corporate Profits to Sales

Table 14 provides detailed, industry-wide information about trends in the profit to salesratio between 1984 and 1988. In the MA group, a decline in the ratio was observed in three of thefive sectors: mining, manufacturing and construction, finance, insurance and real estate. Only inthe latter two sectors was the decline steeper than in the CC and FC groups.

22 The Data

Figure 2Ratio of Corporate Profits to Equity

Graph 2 (and Table 6) show trends in another financial ratio, namely the corporate profitto equity ratio. Canadian-controlled corporations saw their ratio rise slightly, while the ratio forthose under foreign control declined. In the MA group, the profit to equity ratio declined muchmore precipitously. Table 15 shows that a large part of this reduction in the profit to equity ratiowas concentrated in the finance, insurance and real estate sectors. Nevertheless, the other sectorsin the MA group were still relatively weak. The profitability of corporations involved in mergersor acquisitions does not seem to improve rapidly after takeover. This, at least, is the impressionupon examination of two indicators of short-term profitability: the profit to sales ratio and theprofit to equity ratio.

The next two ratios that we examined have more to do with long-term profitability: thecorporate assets to sales ratio and the ratio of corporate assets to sales.

The corporate assets to sales ratio is used as an indicator of investment in physical capitaland is shown in Graph 3 (and Table 7). The ratio increased strongly in all three groups, with thebiggest increase in the FC group. The MA group experienced the weakest increase. However, inview of this group's poor performance in regard to profit, the effort devoted to increasinginvestment in physical capital was the greatest for the group of corporations involved in mergersor acquisitions.

Table 16 shows that the increase in the ratio of corporate assets to sales was quite welldistributed across the industrial sectors. It is interesting to note that, in four of five sectors, theincrease in the MA group's ratio placed it second.

The Data 23

Figure 3Ratio of Corporate Assets to Sales

Figure 4Ratio of Corporate R&D Expenditures to Sales

Finally, Graph 4 and Table 8 show the ratio of research and development expenditures tosales for the various groups. This ratio did not diminish for any group. However, the level ofresearch and development expenditures in the MA group were much stronger initially and alsoincreased much more substantially between 1984 and 1988 than in the other two groups.Corporations targeted in mergers or acquisitions seem to have a much greater tendency to investin research and development even before the transaction takes place. Furthermore, the completion

24 The Data

of the transaction seems to accentuate this difference. Although detailed, industry-wide data arenot available, the increase in the ratio of research and development expenditures to sales is knownto be well distributed among the sectors.

One particular conclusion seems to arise from analysis of these four ratios: the profitabilityof corporations that merged with or were taken over by foreign corporations was not any greaterthan that of the two reference groups of corporations not involved in mergers or acquisitions.However, these corporations did make great strides in "long-term investment" in physical capitaland especially in research and development.

So far, we have drawn comparisons between the group of corporations that were theobjects of mergers or acquisitions by foreign interests and two other groups of corporations thatwere not involved in these activities. Our database did not contain reference groups ofcorporations that merged with or were acquired by Canadian corporations. To cover this gap, wedeveloped Table 9 to show trends in various financial ratios for a group of corporations that weretaken over by Canadian interests. We should point out, first, that the reference period was not thesame as in our sample: the year prior to the acquisitions was 1983 and the year after was 1987. Inaddition, industry representation and company size were very different from our sample.Nevertheless, the comparison is still interesting.

Table 9 shows that the financial returns of corporations that were taken over by Canadianinterests increased considerably after the acquisitions. The profit to sales ratio and the profit toequity ratio increased substantially between 1983 and 1987. On the other hand, investmentindicators show, at best, little change. Assets increased only very slightly in comparison withsales, while research and development expenditures decreased slightly in comparison with sales. Itappears, therefore, that the behaviour of corporations acquired by Canadian interests was exactlythe opposite of the behaviour of corporations acquired by foreign interests.

Corporations acquired by Canadian interests seem to be somewhat myopic in comparisonwith those acquired by foreigners. This different behaviour might however be partly explained bysize differences between Canadian and foreign takeovers in our sample.

The different "vision" between corporations acquired by foreign and Canadian interestscould also be due to differences in the vision of the corporations making the acquisitions. For allsorts of reasons that we will not delve into here, Canadian corporations tend to devote much lessof their budgets to research and development than corporations in other countries. Acquisitions ofCanadian companies by foreign interests may well result in increased research and developmentexpenditures as the targeted companies assume some of the characteristics of foreigncorporations. The Investment Canada working paper, Business Performance Following aTakeover (1992), showed the same differences in profit and research and developmentexpenditures with regard to corporations taken over by Canadian and foreign interests.

25

ECONOMETRIC ANALYSIS

We proceeded with the estimation of the equations described under The ConceptualFramework. The two-stage least-squares method was applied to the system of equations formedby the two individual equations. The observation unit was a vector of variables, as specified in thedata definitions. It was defined for a company on a given date. This means that, for all threegroups, we started with a few more than 770 observations spread over two years, 1984 and 1988.

First we applied a logarithmic transformation to all the variables. We also made acorrection to the initial sample. The distribution of the sample did not follow a normal law: someobservations were at the extremities of the distribution and heavily biased it. Because the purposeof this research was to study the behaviour of corporations that continued to operate normallyafter a takeover, we eliminated two percent of the total number of observations, at both ends ofthe sample, which were causing this bias. The industrial representativeness of the sample remainssimilar after this correction is applied to the initial sample.

Profit to Sales Equation

The estimation results are shown in Table 17. Let us look first at the equation for theprofit to sales ratio. Several variables that were initially included in this equation proved notsignificant, for instance, the unemployment rate variable which was included to capture the effectof the business cycle. However, other variables were already linked closely enough to fluctuationsin the business cycle, and the unemployment rate variable did not add anything to ourmeasurement of cyclical effects.

A dichotomous variable was also added to capture the effect of takeovers on profit tosales ratios. However, this variable proved not significant. This means that the proposition thattakeovers have no short-term effect, whether positive or negative, on profit to sales ratios cannotbe rejected.

Two dichotomous variables did prove significant. The first captured the effect of thecountry in which control was exercised. This variable was "0" if the control was Canadian, and"1" if it was foreign. The results show that Canadian-controlled corporations had profit to salesratios that were generally lower than those of foreign-controlled corporations. However, weobserved earlier that takeovers themselves had no effect on profit. It was therefore not thetakeovers that had a positive effect on corporate profit but rather the fact that control belonged toforeign interests. A certain dynamic seems to emerge in the relationship between profit andcountry of control. Takeovers by foreign interests do not seem to have an immediate positiveimpact on the profitability of the targeted corporations. However, over several years perhaps,foreign-owned corporations do tend to show greater profitability.

26 Econometric Analysis

The second significant dichotomous variable captured the effect of the year when theobservation was made. Its sign was negative. This means that, taking all the explanatory factorsentering the equation into account, the average profit to sales ratio fell between 1984 and 1988. Itis possible that this variable is the one actually capturing the effect of the business cycle on theprofit to sales ratio.

The elasticity of the profit to sales ratio with respect to research and developmentexpenditures was both positive and significant (0.75 on average for the entire sample). This leadsone to think that there is an important connection between a company's research and developmentexpenditures and its profitability. To establish this causal connection, it must be assumed thatcurrent research and development expenditures are a good approximation of past research anddevelopment expenditures. It is, of course, unrealistic to think that research and developmentexpenditures can have a positive impact on profit during the same period. In general, it takesseveral periods before the benefits of research and development expenditures can be realized inthe form of higher profit. If data were available, it would be interesting to replace the currentresearch and development variable with the sum of all expenditures, past and present. Theprobable result would be coefficients showing an increasing, then decreasing importance ofresearch and development expenditures.

The estimated elasticity of the profit to interest rates ratio was negative and significant.This reflects the interpretation of this variable as an approximate measure of the cost of capital.Increases in the cost of capital are reflected in short-term reductions in profitability.

Finally, the elasticity of the profit to sales ratio in comparison with total sales was notsignificant.

The Research and Development Expenditures to Sales Equation

Table 17 also shows the results of the equation for the ratio of a corporation's researchand development expenditures to its total sales. Several variables proved not significant in thiscase as well. For instance, the cyclical unemployment rate variable had no significant effect onprofit rates, nor did the dichotomous variables of the country of control and the year. The ratio ofa corporation's research and development expenditures to sales tends therefore to be rather stableover time.

However, the coefficient of the dichotomous takeover variable was significant. It will beremembered that this variable was "0" if a takeover occurred and "1" otherwise. The variable'ssign of the coefficient meant that corporations that had been taken over experienced, on average,higher investments in research and development than those that had not been taken over. Thisconfirms what was observed earlier about the main characteristics of the data.

The effect of takeovers on research and development expenditures may seem surprising.Most of the other studies we looked at indicated that there was, at best, no relationship betweentakeovers and research and development expenditures. However, we should remember the

Econometric Analysis 27

particular nature of our sample. The corporations were very large and were already under foreigncontrol or passing under it between 1985 and 1987. As was mentioned above, these twopeculiarities may well explain the different behaviour of the corporations in our sample involved inmergers or acquisitions.

The next two coefficients for sales and assets should be interpreted together. The equationfor the ratio of research and development expenditures to sales should theoretically include capitalintensity and not the amount of assets as an independent variable. Since the equation wasestimated using the logarithm of the amount of assets, we must add and subtract 0.05 multipliedby the logarithm of the amount of sales to obtain the equation set forth above. The results showedpositive elasticity for the amount of research and development expenditures in comparison withthe amount of capital. This tends to indicate that the stock of physical capital and investment inresearch and development are two factors that are used in a complementary way in production.

The elasticity of investment in research and development in comparison with total salesremained negative at !0.01 (i.e., !0.06 + 0.05). Total research and development expendituresincreased with rising total sales, but less than proportionally. There were therefore economies ofscale that could be taken advantage of when investing in research and development.

The elasticity of interest rates was positive and significant. This contradicts theinterpretation that interest rates are an approximation of the opportunity cost of the funds devotedto research and development. It is possible that the positive elasticity of interest rates was theresult of a combined effect. The results of Equation 1 indicate, in fact, that if interest rates rise,profit immediately falls. In an attempt to offset this effect on profit, corporations may increasetheir expenditures on research and development. However, it is difficult to confirm thishypothesis, since the profit variable itself was not significant in the research and developmentexpenditure equation.

This econometric analysis confirms the main conclusions of the previous data analysis. Theprofitability of corporations involved in mergers or acquisitions remains, at best, unchanged overthe short run. The profit to sales ratio does not react immediately to takeovers. However,expenditures on research and development are affected positively by takeovers. These resultssupport the hypothesis that the corporations in our sample that were merged or acquired tendedto adopt a more long-term strategy by investing more in physical and technological capital.

29

CONCLUSIONS

From the examination of the data, one observes the following two phenomenon:

• The short-term profitability of corporations is not positively affected by takeovers byforeign interests. The profit-to-sales ratio declined sharply immediately following thetakeovers. The profit to equity ratio also declined.

• Corporations that had been taken over intensified their investments in physical capital orresearch and development.

These observations point to an interesting corporate-adjustment process after takeovers byforeign interests. Immediately after takeovers, corporations adopt a more long-term perspectiveand invest in physical or technological capital. In so doing, they are prepared to accept a short-term decline in profitability.

The results of the econometric analysis do not contradict these hypotheses. Indeed, wefound that:

• The coefficients of the first equation, the profit to sales ratio, indicate that takeovers donot appear to have any immediate effect on corporate profitability. The post takeoverdecline for the MA group can be mostly explained by cyclical factors.

• The country of control of the corporation, whether it is Canadian or foreign, is having asignificant impact on profitability, to the advantage of the foreign-controlled corporation.

It is therefore possible that, at the time when foreign interests take control of a Canadiancorporation, profitability remains relatively unchanged or even declines. However, in the longerrun, as the new "corporate culture" transfers to the newly acquired corporation, profitabilityincreases and eventually surpasses the average for all Canadian-controlled corporations. Thisprocess of transferring the corporate culture can take several forms, including investment inhuman resources, such as training, and in technology.

• The equation for the ratio of research and development expenditures to sales confirms thatinvestment in research and development increases after such takeovers.

We should recall here another result obtained when comparing our sample with that ofanother study which dealt with corporations taken over by Canadian interests. In contrast totakeovers by foreign interests, takeovers by Canadian interests result in higher profitability andlower investment in capital and research and development in the targeted corporations afteracquisition. These results seem to indicate that there is a different "culture" in foreign- andCanadian-owned corporations.

30 Conclusions

This research shows, in our view, the importance of analyzing the consequences ofmergers and acquisitions over time. The low profitability of the targeted corporations in the firstmonths and years after the takeovers would be the result of a period of adjustment during whichthe two working cultures can get to know each other and learn how to co-operate. In addition,the corporation taking over has to take time to get to know and use profitably all of the variousassets it has acquired.

Corporate assets are increasingly intangible. The internalization and technologicalcompetence theories outlined in this paper suggest that a desire to acquire intangible assets maybe an important motive in acquisitions.

In order for corporations to profit fully from these intangible assets, they need time andinvestments, often complementary, in physical capital and research and development. The researchresults obtained here are certainly compatible with our interpretation of these internalizationtheories and the evolving adjustment that necessarily results.

The results of our research support even greater liberalization of foreign investment. Thecompetitiveness of the entire Canadian economy improves if the long-term profitability ofparticular corporations increases as a result of the intervention of foreign interests which are ableto make investments thanks to their greater size, their financial resources, their patience and morelong-term perspective, their complementary assets and the synergies that result.

Future research in evaluating the adjustment of firms to takeover could be built on someof the weaknesses already identified in this research. It would be helpful, in particular, to have tohave the database augmented with several additional years of observations before and after thetakeovers. If it is true that the acquiring corporation is ready to accept short-term declines inprofitability in order for investments to be made and long-term profitability increased, then thedata should enable us to observe this adjustment process.

The relatively limited number of takeovers left us with a sample whose distribution did notreflect the norm. It was therefore difficult in some cases to draw clear, unequivocal conclusions. Itis essential to increase the number of takeovers in our initial sample. This would make it possibleto improve the representativeness of the sample and extend more easily the period of analysisbefore and after the takeovers.

In addition, the limited number of variables available did not enable us to analyze all theaspects of the adjustment process that corporations go through after being taken over. It wouldbe helpful to increase the number of variables on the economic behaviour of corporations (toinclude for instance the number of employees, training data and productivity data) in order to beable to analyze all the aspects of the adjustment process more effectively.

31

APPENDIX ADESCRIPTION AND LIMITS OF THE DATA SET

Description of the Data Set

A list of foreign acquisitions of Canadian corporations approved by Investment Canadabetween July 1, 1985 and December 31, 1987 was provided to Statistics Canada. This list waschecked using information collected under the Corporations and Labour Unions Returns Act(CALURA).

The information from the CALURA data bank showed that, in 11 cases, the approvedacquisition never took place and in 127 cases, it pertained to only part of the legal entity makingup the corporation. In all cases, the original legal entity continued to exist and was not absorbedby the foreign purchaser. In 115 cases, the acquired corporation was merged into the foreign-owned corporation and ceased to exist as a legal entity. In addition, two corporations were takenover twice during the three-year period of the study. In 131 cases, the legal entity that was takenover continued to exist as a legal entity after the acquisition. These 131 corporations were thegroup that we studied.

Financial Characteristics of the Corporations

For the 131 corporations in the study, a connection was established with the governmenttax files for the years 1984 and 1988. The tax data provide only a limited view of the mainfinancial characteristics of all corporations active in Canada. Four characteristics were pulled outfor the years 1984 and 1988.

Main Financial Characteristics

Assets

These are the total assets on corporate balance sheets. They include such items as cashbalances, exchangeable securities, accounts payable, inventory, net fixed assets, investments inaffiliated corporations and other kinds of assets.

Equity

This reflects stockholder participation in the net assets of a corporation. It generallyincludes the total amount of issued and paid-up share capital, undistributed profit and othersurplus accounts, such as initial and capital surpluses.

32 Appendix A

Sales

In the case of corporations in non-financial sectors, "sales" represents gross revenues fromnon-financial operations. In the finance, insurance and real estate sector, "sales" represents totalrevenues from both financial and non-financial sources.

Profit

The net value of all book profits and losses. This includes extraordinary items. The amountof profit is determined before deduction of income taxes.

Research and Development

The 131 corporations were connected to the research and development database kept bythe Services, Science and Technology Division of Statistics Canada.

Expenditures on research and development were collected as part of a Statistics Canadastudy and were looked at for 1984 and 1988.

Expenditures were considered to be on research and development if they pertained to: • systematic research in the natural sciences or engineering based on experiments or

analysis and conducted for scientific or commercial progress;• original research conducted in a systematic way in order to acquire new knowledge; or • the application of research results or other scientific knowledge to creating new or

substantially improved products or processes.

In general, expenditures were considered to be on research and development if theycomplied with the definition of "scientific research and experimental development" in the CanadaIncome Tax Act. For a more detailed description of research and development, see Statistics onIndustrial Research and Development, publication no. 88-202 of the Statistics Canada catalogue.

Industry Classification

The corporations in this study were classified according to the Standard IndustrialClassification (SIC) of 1960, in their entirety in particular industries, according to the activitiesthat produced most of their gross revenues, even though the large corporations may have beeninvolved in several different industrial categories. As a result of reorganizations, corporationsmight be reclassified into other industrial categories. To maintain the integrity of our analysis, weused the SIC corresponding to the 1984 data. We were forced to

Appendix A 33

amalgamate the industries into large groups for reasons of data confidentiality. The large groupsor industrial sectors used in this study were:

• agriculture, forestry, fishing and mining;• manufacturing and construction;• the wholesale and retail trade;• finance, insurance and real estate; and• services and public amenities.

Selection of Reference Groups

The study attempted to determine if the acquisitions approved by Investment Canada hadany particular characteristics in comparison with all other corporations. However, the 131selected corporations were not representative of all corporations from the points of view of size,industrial distribution, longevity or rate of return. As a result, they could not be compared withall other industries. In addition, the selected corporations were themselves quite disparate in sizeand rate of return, and the variables diverged from the norm in ways that were not in keeping withnormal distribution. The deviations were biased, with large tails. For these reasons, we decidedthat the comparative analysis would be based on matched samples and non-parametric techniques.

The reference groups were designed to be as similar as possible to the group ofcorporations that had been taken over. The two matched reference groups had the followingcharacteristics:

• The matched corporations were active throughout the entire five years covered by thestudy (1984 to 1988).

• The matched corporations had not been the objects of acquisitions, transfers or takeoversduring the period in question.

• Crown corporations were excluded.

• The matched corporations had the same SIC (at a three-figure level of aggregation) in1984 as the corresponding corporations that had been taken over.

• The matched corporations belonged to the same asset cohort in 1984 as the corporationsthat had been taken over. These asset cohorts were:

• less than $1 million;• from $1 million to $25 million;• from $25 million to $50 million; • from $50 million to $100 million; and • more than $100 million.

34 Appendix A

• Two matched groups were established: one comprised solely Canadian-controlledcorporations and another comprised solely foreign-controlled corporations.

• If more than one corporation fit the characteristics outlined above, the matchedcorporation was chosen at random among those that fit.

Warning

Analysis showed that around 70 percent of the 131 corporations registered an absolutevariation of more than 100 percent in one or another of their financial variables over the periodunder study. In addition, the direction of the changes was not uniform. Some of thesecorporations were substantially restructured after being acquired, while others benefited fromfresh injections of capital. Still others had some of their assets assigned or transferred tosubsidiaries.

A ' P × Y & w × L & rk × K & rrd × RD

L ' L ( p , w , rk , rrd ) w ' w ( L , K , RD , Y )

K ' K ( p , w , rk , rrd ) rk ' rk ( L , K , RD , Y )RD ' RD ( p , w , rk , rrd ) rrd ' rrd ( L , K , RD , Y )

Y ' Y ( p , w , rk , rrd ) p ' p ( L , K , RD , Y )

35

APPENDIX BANALYSIS OF THE PROFIT FUNCTION

Corporate profit was defined as the difference between revenues from the sale of goodsand the cost of producing these goods. The profit identity looked as follows:

(3)

where II represents profitP the cost of productionY production

L, K and RD purchases of the labour, physical capital and research and development inputsw, r and r the cost of labour, capital and research and development.k rd

Since it is an identity, this expression of profit can not be estimated. We used it instead toderive the profit equation containing the parameters to be determined. Let us assume the existenceof a production function relating inputs to outputs and possessing normal continuity andconcavity. Solving the problem of profit maximization under the constraints of the productionfunction makes it possible to obtain functions for product supply and demand for the followingfactors:

The four equations on the left-hand side are the supply and demand functions expressed invalue space. This is the way these equations are usually presented. The four equations on theright-hand side are duals of the equations on the right. They express the supply and demandfunctions in quantity space. They can be considered marginal evaluation functions of the quantitiessupplied and demanded.

The profit maximization function can be obtained by substituting, in the profit identity, thefunctions of supply and demand—or of marginal evaluation—shown above. It is thereforepossible to obtain a profit function expressed in value space, i.e., prices, or in quantity space. It isalso possible, by substituting sometimes one and then the other, to obtain a mixed profit function.This is what we decided to do. The profit equation in this study came from substituting, in the

36 Appendix B

profit identity, capital demand and product supply equations and marginal evaluation functions oflabour and research and development.

37

BIBLIOGRAPHY

BALDWIN, J.R. and R.E. CAVES. "Foreign Multinational Enterprises and Merger Activity inCanada," paper presented at the Investment Canada conference, Corporate Globalizationthrough Mergers and Acquisitions, Toronto, November 1990.

BALDWIN, J.R. and P.K. GORECKI. Mergers and Merger Policy in the CanadianManufacturing Sector: 1971-1979, Working Paper No. 297, Economic Council ofCanada, Ottawa, 1986.

BALDWIN, J.R. and P.K. GORECKI. "Plant Creation versus Plant Acquisition: The EntryProcess in Canadian Manufacturing," International Journal of Industrial Organization 5,March 1987, pp. 27-41.

BROWN C. and J.L. MEDOFF, "The Impact of Firm Acquisition on Labor," CorporateTakeovers: Causes and Consequences, published under Alan Auerbach, University ofChicago Press, Chicago, 1988, pp. 9-25.

CAVES, R.E. "Mergers, Takeovers and Economic Efficiency: Foresight vs. Hindsight,"International Journal of Industrial Organization 7, 1989, pp. 151-174.

HALL, B.H. "The Effect of Takeover Activity on Corporate Research and Development,"Corporate Takeovers: Causes and Consequences published under Alan Auerbach,University of Chicago Press, Chicago, 1988, pp. 69-100.

INVESTMENT CANADA. Business Performance Following a Takeover, Working PaperNo. 11, April 1992.

INVESTMENT CANADA. Issues Surrounding Merger and Acquisition Activity in Canada andby Canadian Firms Abroad, Working Paper No. 13, March 1993.

JENSEN, M.C. "Agency Costs of Free Cash Flow, Corporate Finance and Takeovers," AmericanEconomic Review 76, no. 2, May 1986.

KHEMANI, R.S. "Recent Trends in Merger and Acquisition Activity in Canada and SelectedCountries," paper presented at the Investment Canada conference, CorporateGlobalization through Mergers and Acquisitions, Toronto, November 1990.

LICHTENBERG, F.R. Corporate Takeovers and Productivity, MIT Press, Cambridge, MA,1992.

38 Bibliography

RAVENSCRAFT D.J. and F.M. SCHERER. Mergers, Sell-Offs and Economic Efficiency,Washington, D.C., Brookings Institution, 1987.

RAVENSCRAFT D.J. and F.M. SCHERER. "The Profitability of Mergers," InternationalJournal of Industrial Organization 7, 1989.

TARASOFSKY A. and R. CORVARI. Corporate Mergers and Acquisitions: Evidence onProfitability, study produced by the Economic Council of Canada, Ottawa, 1991.

TABLES

Tables 41

TABLE 1Average corporate revenues, by group and country of control, 1984 and 1988(in thousands of dollars)

1984 1988 ª%

MA Group 120,870 131,314 8.6

CC Group 98,354 129,739 31.9

FC Group 123,843 138,931 12.2

Total, Canadian Control 1,359 1,454 7.0

Total, Foreign Control 47,761 50,498 5.7

TABLE 2Average corporate profit, by group and country of control, 1984 and 1988(in thousands of dollars)

1984 1988 ª%

MA Group 8,118 8,172 0.7

CC Group 8,371 11,072 32.3

FC Group 12,764 11,043 -13.5

Total, Canadian Control 67 88 31.3

Total, Foreign Control 4, 214 3,529 -16.3

TABLE 3Average corporate assets, by group and country of control, 1984 and 1988(in thousands of dollars)

1984 1988 ª%

MA Group 133,750 174,155 30.2

CC Group 126,377 161,251 27.6

FC Group 106,542 137,910 29.4

Total, Canadian Control 1,287 1,317 2.3

42 Tables

Total, Foreign Control 34,211 44,015 28.7

TABLE 4Average corporate equity, by group and country of control, 1984 and 1988(in thousands of dollars)

1984 1988 ª%

MA Group 55,621 69,553 25.0

CC Group 62,388 88,806 42.3

FC Group 53,761 64,283 19.6

Total, Canadian Control 403 461 14.4

Total, Foreign Control 15,995 19,421 21.4

TABLE 5Ratio of corporate profit to sales, by group and country of control, 1984 and 1988

1984 1988 ª

MA Group 0.17 -0.57 -0.74

CC Group 0.13 -0.49 -0.62

FC Group 0.13 0.02 -0.11

Total, Canadian Control 0.04 0.06 0.02

Total, Foreign Control 0.08 0.07 -0.01

TABLE 6Ratio of corporate profit to equity, by group and country of control, 1984 and 1988

1984 1988 ª

MA Group 0.18 -3.10 -3.28

CC Group 0.28 0.35 0.07

FC Group 0.35 0.22 -0.13

Total, Canadian Control 0.16 0.18 0.02

Total, Foreign Control 0.26 0.19 -0.07

Tables 43

TABLE 7Ratio of corporate assets to sales, by group and country of control, 1984 and 1988

1984 1988 ª

MA Group 2.54 4.23 1.69

CC Group 3.66 5.75 2.09

FC Group 2.37 5.16 2.79

Total, Canadian Control 0.95 0.87 -0.08

Total, Foreign Control 0.73 0.77 0.04

TABLE 8Ratio of corporate R&D expenditures to sales, by group and country of control, 1984 and1988

1984 1988 ª

MA Group 0.15 0.43 0.28

CC Group 0.03 0.04 0.01

FC Group 0.04 0.04 0.00

Total, Canadian Control 0.015 0.016 0.001

Total, Foreign Control 0.009 0.012 0.003

TABLE 9Financial ratios of corporations that were the objects of mergers or acquisitions byCanadian corporations, 1983 and 1987

1983 1987 ª

Ratio of Profit to Sales 0.02 0.06 0.04

Ratio of Profit to Equity 0.07 0.21 0.14

Ratio of Assets to Sales 0.72 0.77 0.05

Ratio of R&D Expenditures to Sales 0.0035 0.0026 -0.0009

TABLE 10Average corporate revenues, by group and country of control, 1984 and 1988, according to industrial group(in millions of dollars)

Mining Manufacturing Wholesale and Finance, Insurance Services and Construction Retail Trade and Real Estate and Public Amenities

1984 1988 ª% 198 198 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª%4 8

MA Group 205 103 -49.8 99 116 17.1 212 235 10.9 30 91 203. 30 26 -13.37 3

CC Group 60 37 -38.0 89 121 35.7 175 238 35.9 27 46 71.5 89 110 23.1

FC Group 260 159 -38.9 104 134 29.0 202 240 18.4 14 17 27.7 29 39 33.9

Total, Canadian Control 3 3 - 1.7 2 17.6 1.6 2 25.0 - - - 0.9 0.8 -11.2

Total, Foreign Control 59 33 -44.1 71.3 80.2 12.5 31.8 43 35.2 - - - 15.5 13.0 -16.1

TABLE 11Average corporate profit, by group and country of control, 1984 and 1988, according to industrial group(in millions of dollars)

Mining Manufacturing Wholesale Finance, Insurance Services and Construction and Retail Trade and Real Estate and Public Amenities

1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª%

MA Group 50 -0.3 -100.6 3.9 10.6 171.8 2.4 4.3 -44.19 7.6 15.9 109.2 2.5 1.8 -28.0

CC Group 11 1.7 -84.5 7.1 10.1 42.3 1.5 10.6 606.7 3.4 6.9 102.9 41.0 35.7 -12.9

FC Group 78 16 -79.5 8.5 13.5 58.8 4.2 7.0 66.6 1.4 8.1 478.6 5.1 6.0 17.6

Total, CanadianControl

0.4 0.5 25.0 0.07 0.01 42.9 0.06 0.06 - - - - 0.07 0.07 -

Total, ForeignControl

18 3 -83.3 5.2 6.8 30.8 0.75 1.5 50.0 - - - 2.0 1.1 -45.0

TABLE 12Average corporate assets, by group and country of control, 1984 and 1988, according to industrial group(in millions of dollars)

Mining Manufacturing Wholesale Finance, Insurance Services and Construction and Retail Trade and Real Estate and Public Amenities

1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª%

MA Group 621 713 14.8 75 101 34.70 70 76 8.60 179 316 76.50 47 73 55.30

CC Group 233 188 -19.3 86 129 51 111 150 35.1 88 154 75.9 367 381 3.9

FC Group 383 575 49.9 79 97 23.3 93 116 25.2 76 57 -25 37 64 73.7

Total, Canadian Control 11.0 9.0 -18.2 1.3 1.5 15.4 0.7 0.8 14.3 - - - 1.5 1.4 -6.7

Total, Foreign Control 103 129 25.2 46 60 30.4 13 22 69.2 - - - 15.9 17.2 8.2

TABLE 13Average corporate equity, by group and country of control, 1984 and 1988, according to industrial group(in millions of dollars)

Mining Manufacturing Wholesale Finance, Insurance Services and Construction and Retail Trade and Real Estate and Public Amenities

1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª% 1984 1988 ª%

MA Group 221 274 23.6 35 49 40.00 41 35 -14.60 56 72 28.60 18 28 55.60

CC Group 111 114 3.1 47 63 34.7 45 74 65.3 31 49 60.4 204 343 68

FC Group 228 245 7.5 40 53 34.1 38 47 22.2 34 22 -33.3 16 38 139

Total, CanadianControl

5.0 4.0 -20 0.5 0.6 20 0.2 0.2 - - - - 0.4 0.4 -

Total, Foreign Control 47 57 21.3 23 29 26.1 4.5 8 77.8 - - - 6.2 6.2 -

TABLE 14Ratio of corporate profit to sales, by group and country of control, 1984 and 1988, according to industrial group

Mining Manufacturing Wholesale Finance, Insurance Services and Construction and Retail Trade and Real Estate and Public Amenities

1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª

MA Group 0.27 -3.01 - 0.10 -0.25 -0.35 0.02 0.04 0.02 0.75 -1.70 - 0.07 0.14 0.073.28 2.4

5

CC Group 0.31 -5.85 - 0.06 0.00 -0.06 0.16 0.06 -0.10 0.26 0.25 - 0.08 0.04 -0.046.16 0.0

1

FC Group 0.28 0.04 - 0.08 0.09 0.01 0.07 -0.08 -0.15 0.21 1.01 0.8 0.28 0.24 -0.040.24 0

Total, Canadian Control 0.11 0.17 0.06 0.04 0.07 0.03 0.03 0.03 - 0.13 0.17 0.0 0.08 0.09 0.014

Total, Foreign Control 0.31 0.10 - 0.07 0.08 0.01 0.02 0.03 0.01 0.12 0.20 0.0 0.13 0.09 -0.040.21 8

TABLE 15Ratio of corporate profit to equity, by group and country of control, 1984 and 1988, according to industrial group

Mining Manufacturing and Wholesale and Retail Finance, Insurance and Services and PublicConstruction Trade Real Estate Amenities

1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª

MA Group -0.05 -0.09 - 0.18 0.22 0.04 0.19 0.21 0.02 0.25 - - 0.34 0.13 -0.04 27.86 28.1 0.21

1

CC Group 0.09 -0.22 - 0.41 0.21 - 0.14 0.91 0.77 0.18 0.39 0.21 0.29 0.15 -0.31 0.20 0.14

FC Group 0.11 -0.71 - 0.23 0.21 - 0.46 0.39 s- 0.19 0.46 0.27 1.75 0.70 -0.82 0.02 0.07 1.05

Total, Canadian Control 0.08 0.10 0.02 0.15 0.22 0.07 0.25 0.27 0.02 - - - 0.17 0.17 -

Total, Foreign Control 0.38 0.06 - 0.22 0.23 0.01 0.17 0.18 0.01 - - - 0.32 0.18 -0.32 0.14

TABLE 16Ratio of corporate assets to sales, by group and country of control, 1984 and 1988, according to industrial group

Mining Manufacturing and Wholesale and Retail Finance, Insurance and Services and PublicConstruction Trade Real Estate Amenities

1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª 1984 1988 ª

MA Group 3.40 9.68 6.28 0.85 2.29 1.44 0.46 1.25 0.79 13.9 15.0 1.14 1.43 2.70 1.273 7

CC Group 22.3 21.3 -0.98 1.05 3.16 2.11 0.59 6.63 6.04 7.01 5.57 -1.44 2.11 1.59 -0.520 2

FC Group 4.35 14.9 10.57 0.96 1.24 0.28 1.14 2.94 1.80 9.44 18.2 8.80 1.31 4.35 3.042 4

Total, Canadian Control 3.26 3.13 -0.13 0.76 0.80 0.04 0.42 0.40 -0.02 7.09 7.56 0.47 1.70 1.67 -0.03

Total, Foreign Control 1.75 3.88 2.13 0.64 0.75 0.11 0.40 0.51 -0.11 5.08 6.06 0.98 1.03 1.35 0.32

48

TABLE 17Results of estimations, two simultaneous equations1

Equation for the Equation for the Profit to Sales Ratio R&D Expenditures to Sales Ratio

Parameter Statistic t Parameter Statistic t

Constant -1.85 (-2.62) 0.19 (1.84)

Sales -0.04 (-0.58) -0.06 (-4.85)

Assets - - 0.05 (3.37)

Real Interest Rates -0.14 (-4.69) 0.01 (2.33)

R&D Expenditures to Sales Ratio 6.21 (4.03) - -

Dichotomous Variable, Country of Control(foreign control = 1)

0.36 (1.82) - -

Dichotomous Variable, Takeover(takeover = 1)

- - 0.13 (2.98)

Dichotomous Variable, Year(1988 = 1) -0.56 (-4.03) -0.03 (-0.74)

Overall Statistics:NR2

Value of F

574 5740.06 0.058.68 7.17

All variables are expressed as natural logarithms except the R&D expenditures to sales ratio which is1

expressed as a level.